PC Random Shutdown at Gaming

[tolagarf]

I didn't really know where to place this, since I'm not really sure what is the cause of my issue here. I'm leaning towards perhaps a driver issue with the latest nvidia drivers, since they seem to be rather faulty (362.00 / 364.51). Hopefully someone's got an idea.

So I've started getting random shutdowns (reboot) while gaming, which has happened a couple of times playing The Division and GTA V. It just happens suddenly, but not very often. The PC loses power and turns itself on again. Looking at the event viewer it will say something like the shutdown at 18:41 was unexpected and gives a Kernen-Power error, however looking through the logs there is nothing at all at that time period. Actually logs from several hours seem to be missing.

So far I've tried defaulting all my OC, everything is factory default right now. Doesn't seem to help much. I'm currently in the process of switching back to nvidia driver version 361.91 and hope it solves itself. So I'm leaning towards either PSU is dying (not again please!), nvidia driver crash, or perhaps even my Predator pump sending the shutdown signal due to some unforeseen issue (It's not overheating though under stress).

So to sum it up: Never happens during stress testing with Prime95 or OCCT (done over an hour testing). Only happens while gaming, seems to be rare occurrence.

System specs:

Asus Sabertooth X99 m/b
Corsair DDR4 4 x 4 GB 2800 MHz RAM CL16
i7 5930K CPU
Asus Poseidon GTX 980 Ti
Corsair AX860i PSU
Samsung SSD 850 Pro 512GB
EKWB Predator XLC-360 AIO

 

[Damien c]

Personally I would run something like Heaven or 3D Mark etc to stress the GPU to check that and then I would run the same program with OCCT or Prime95 at the same as the GPU stress test, if that all past I would run Memtest.

Have you got another PSU or GPU etc that you can test in case it fails when doing all that?

 

[tolagarf]

Personally I would run something like Heaven or 3D Mark etc to stress the GPU to check that and then I would run the same program with OCCT or Prime95 at the same as the GPU stress test, if that all past I would run Memtest.

Have you got another PSU or GPU etc that you can test in case it fails when doing all that?

Guess I could set Valley benchmark to run over night, but 3D Mark I can't loop as one of the benchmarks crashes the PC and gives me an overclock fail error when rebooting. Think it's the one with the flying eagle and the upside down temple, and it always happens in the exact same spot which makes me think it's actually a bug in 3D Mark.

I've also cleaned out the 364.xx driver with DDU and reinstalled 361.91, so I'm hoping this is what's giving me grief...

 

[Damien c]

Does that happen on all 3D Mark suites or a specific one?

I was thinking possibly the ram speed could be a issue and try lowering it but as you say it passes OCCT and Prime for over an hour, then really that would show a ram issue.

Would be worth running Heaven or something like that and see if it happens.

 

[tolagarf]

Does that happen on all 3D Mark suites or a specific one?

I was thinking possibly the ram speed could be a issue and try lowering it but as you say it passes OCCT and Prime for over an hour, then really that would show a ram issue.

Would be worth running Heaven or something like that and see if it happens.

Ran Valley for 9 hours straight without an issue. Actually the graphics card is making less buzzing noise now when rendering. And heaven benchmark is just an older version compared to Valley. Both are Unigine benchmarks. RAM is currently running at 2133 MHz, next up I'll try and bring them up to 2400 again and do a thorough memtest.

Regarding 3D Mark it happens in the only version I have (brought on Steam). Ehm I'm not really sure which version it is except it's from 2014/2015. I did check up on that crash problem some time ago, it seem several people have that issue who owns an i7 5930K CPU. Can't see why it would be specific to that CPU only though, but seems so.

Edit: Mem test at 2400 MHz CL15 1.25v successful. No errors.